Theor Appl Genet (1990) 80:281-287
9 Springer-Verlag 1990
Dosage effects of chromosomes of homoeologous groups 1 and 6 upon bread-making quality in hexaploid wheat W.J. Rogers, J.M. Riekatson, E.J. Sayers and C.N. Law Cambridge Laboratory, JI Centre for Plant Science Research, Colney Lane, Norwich NR4 7UJ, UK Received January 12, 1990; Accepted February 23, 1990 Communicated by J.W. Snape
Summary. The endosperm storage proteins, glutenin and gliadin, are major determinants of bread-making quality in hexaploid wheat. Genes encoding them are located on chromosomes of homoeologous groups 1 and 6. Aneuploid lines of these groups in spring wheat cultivar 'Chinese Spring' have been used to investigate the effect of varying the dosage of chromosomes and chromosome arms upon bread-making quality, where quality has been assessed using the SDS-sedimentation test. Differences between the group 1 chromosomes for quality were greater than those between the group 6 chromosomes. The chromosomes were ranked within homoeologous groups for their effect on quality as follows ( > = b e t t e r quality): 1 D > I B > I A and 6 A > 6 B = 6 D . The relationship of chromosome dosage with quality was principally linear for four of the chromosomes, but not for 6B and 6D. Increases in the dosage of 1B, 6A and, especially, 1D, were associated with significant improvements in quality, whereas increases in the dosage of IA were associated with reductions in quality. The effects of 1A and 1D were such that the best genotype for quality was nullisomic 1A-tetrasomic 1D. For group 1, effects of the long arm appeared in general to be more important than effects of the short arm. For group 6, effects were found associated with the long arms as well as with the short arms, a surprising result in view of the absence of genes encoding storage proteins on the long arms. Significant interactions were found between chromosomes and genetic backgrounds, and between individual chromosomes. Analysis of trials grown over two years demonstrated that, although additive environmental differences over years and genotype x years interaction were present, they were relatively small in magnitude compared with purely genetic differences.
Key words: Triticum aestivum - Glutenin - Gliadin Aneuploid
Introduction Most studies of the relationship between endosperm storage proteins and bread-making quality in hexaploid wheat Triticum aestivum L. have sought to characterise the effect of allelic variation in genetic lines or commercial cultivars (Sozinov and Poperelya 1980; Payne et al. 1984, 1987a). The result has been the development of valuable methods for obtaining improvements in quality, by the identification of alleles associated with high quality, and their consequent assembly into favourable combinations during a breeding programme (Payne et al. 1984). Less attention has been directed towards the influence on quality of varying the dosage of genes known to affect quality. The principal series of homoeoloci encoding storage proteins that govern bread-making quality in hexaploid wheat are located on chromosomes comprising homoeologous groups 1 and 6 (Payne 1987, for review): Glu-1, located on the long arms of chromosomes of group 1, encoding high-molecular-weight (HMW) subunits of glutenin; Gli-1, located on the short arms of chromosomes of group 1, encoding co-gliadins, most of the 7gliadins and some of the fl-gliadins; Gli-2, located on the short arms of chromosomes of group 6, encoding egliadins, most of the fl-gliadins and some of the 7gliadins; and Glu-3, tightly linked to GIi-1 on group 1, encoding low-molecular-weight (LMW) subunits of glutenin. Several additional loci encoding endosperm proteins are located on the short arms of group 1, allelic variation at which has not yet been shown to influence quality. These loci include Gli-3, encoding o~-type gliadins and D subunits of LMW glutenin (Payne et al. 1988); loci encoding further minor classes of gliadins (Akhmedov and Metakovsky 1987); and Tri-l, encoding triticins that are classified as globulins (Singh et al. 1988) rather than as prolamins.
282 Studies of dosage have been primarily concerned with the effect of reduction in the dosage or, indeed, of complete removal, of particular loci, chromosome segments or whole chromosomes. F r o m such studies it has emerged that the following p h e n o m e n a detrimentally affect quality: (i) replacement of certain active alleles by null alleles, particularly at Glu-D1 (Payne et al. 1987b; Lawrence et al. 1988); (ii) replacement of part or all of the short arm of 1D and, to a lesser extent, of 1B, carrying Gli-1, by part of the short arms of homoeologous chromosomes from rye (Koebner et al. 1984) and Aegilops umbellulata (Harris 1983; Rogers et al. 1987); (iii) complete removal of the short arm of 1D in a ditelosomic line (Maystrenko etal. 1973); a n d (iv) reduction of the dosage of chromosome ID to the monosomic level (Welsh and H e h n 1964). Also, the replacement of Gli-B3 by a segment of chromosome 1U from Ae. umbeIlulata has been implicated in lowering quality (Harris 1983; Rogers et al. 1987). The current paper assesses the effect of varying the dosage of whole chromosomes and chromosome arms, both in the decreasing direction, i.e. below the disomic level, as in the previous studies cited above, a n d in the increasing direction, above the disomic level, in an attempt to identify beneficial effects on quality that might be open to exploitation via cytogenetic m a n i p u l a t i o n or molecular biology.
Materials and methods The study was principally based upon a trial of aneuploid lines of spring wheat cultivar 'Chinese Spring' (CS) grown during 1988 (trial T88). The CS aneuploids were originally produced by Dr. E. R. Sears, University of Missouri, and seed was kindly provided from stocks maintained by T. E. Miller and S. M. Reader at the Cambridge Laboratory. T88 consisted of 32 aneuploid lines involving chromosomes from homoeologous groups 1 and 6, as shown in Table 1, plus CS euploid controls. The abbreviations for the aneuploids given in Table I (NT, T, etc) will be used throughout the text. Six samples of CS were grown, where three were arbitrarily assigned to each group of aneuploids (1 and 6), enabling a number of a priori comparisons to be carried out during analysis (see 'Resuits'). A randomised complete block design was used, where each line was represented by a single plot of six plants in each of five independently randomised replicate blocks. Individual seeds were initially germinated on moist filter paper. Roots were taken from seedlings of monosomic families for cytological checking of Feulgen-stained mitotic preparations, and individuals with other than 41 chromosomes were discarded. Seedlings were planted into 4-cm pots and subsequently transplanted into soil in five unheated irrigated glasshouse compartments. Spacing was 15 cm between plants within a row (plot) and 20 cm between rows. Guard rows were sown at the ends of replicate blocks to prevent edge effects. During growth, only one fertiliser application was made, shortly after ear emergence, to guard against artifically boosting the protein content, which might obscure differences in quality due to differences in protein composition. Individual ears were bagged prior to anthesis to guar-
Table 1. Genotypes grown in trial T88 a Group 1 Nullisomic-tetrasomic (NT)
NtA-T1B NIA-TID
N1B-T1A NIB-T1D
N1D-T1A NID-T1B
Tetrasomic (T)
T1A
TIB
TID
Monosomic (M)
MIA
M1B
M1D
Ditelosomic (DT)
DT1AS DT1AL
DT1BS DTIBL
DT1DL
'Double ditelosomic' (DDT)
-
-
DDT1D b
Group 6 Nullisomic-tetrasomic (NT)
N6A-T6B N6A-T6D
N6B-T6A -
N6D-T6A N6D-T6B
Tetrasomic (T)
T6A
T6B
T6D
Monosomic (M)
M6A
M6B
M6D
Ditelosomic (DT)
-
DT6BS -
DT6DS DT6DL
Plus six independent euploid controls b This line is not a true double ditelosomic; one telosomic pair represents chromosome arm IDL. The other is not a true telosome; instead, it appears to be deleted long arm separated by the centromere from, presumably, a small segment of the short arm
antee self-pollination. Harvesting was carried out at maturity, and grain yield (GY) data was obtained for individual plants.
Quality tests Shortly after harvesting, seed was milled using a Tecator Cyclotec sample mill, and the flour was conditioned to 14%-15% moisture content. Grain protein contents (PC) based on 14% moisture content were determined using near infrared reflectance spectroscopy. Due to the very low GY from some of the aneuploids, it was necessary to apply a reliable quality test that could be performed on small amounts of material. For this purpose, a procedure was adapted from a test used to screen early generation durum wheat breeding lines, for gluten strength, using 1-g flour samples (Dick and Quick 1983), and termed the SDS-microsedimentation test (MST). Preliminary tests, in which the volume of a stock solution of SDS/lactic acid and weight of flour samples were varied, suggested that 0.7 g samples, 4 ml of HzO and 12 ml of stock solution would give the greatest differentiation between MST values for this material. The results were expressed as height in cm of sediment formed in a glass test tube (150 mm long x 16mm OD x 14ram ID) after 30min. Two replicate MST determinations were made on each flour sample.
Results
Correlation between characters N o correlation was found between PC a n d MST, suggesting that overall differences between genotypes for M S T were due to qualitative differences in protein composition rather than to quantitative differences in protein
283 amount. To ensure against misinterpretation, M S T was analysed b o t h with and without PC as a covariate, and the conclusions differed appreciably in only one case (considered below). In general, only analyses unadjusted for PC are reported. Individual chromosome effects The mean performance o f each genotype is given in Table 2. F o r analysis, the NTs and Ts of each homoeologous group, plus three o f the CS euploids, provide a balanced and orthogonal set o f genotypes for identifying the effects of individual chromosomes and combinations of chromosomes, in an extension of the design of Pink and Law (1985). In theory, with the families arranged as shown in Table 2, each c h r o m o s o m e appears in four ' b a c k g r o u n d s ' , i.e. the three pairwise combinations of the remaining two chromosomes in the relevant homoeologous group, plus the combination of all three chromosomes. F o r each chromosome, the mean over the four b a c k g r o u n d s can be c o m p a r e d to determine which chrom o s o m e is best for quality; similarly, the means of the b a c k g r o u n d s can be compared. Interactions between chromosomes can also be tested. In practice, b a c k g r o u n d 6A-6D has h a d to be excluded due to the enforced absence of N 6 B - T 6 D from the trial (Table 1). A n analysis o f variance of the d a t a is given in Table 3, including orthogonal comparisons between chromosomes. Previous evidence suggested that valid comparisons would be 1D against the mean o f 1A and 1B, and, given that 1A carries a null allele at Glu-A1, IB-1D against the mean o f 1A-1B and 1A-1D. Also the euploid b a c k g r o u n d can be c o m p a r e d with the m e a n o f the three pairwise comparisons 1A-1B, 1A-1D and 1B-1D. N o a priori comparisons have been calculated for group 6, but differences can be c o m p a r e d using the a p p r o p r i a t e stand a r d error of the difference between means. The analyses clearly demonstrate the superiority of 1D over the mean of 1A and 1B, and o f 1B over I A . The best b a c k g r o u n d is 1B-1D, and 1A-1D is superior to 1A-lB. F o r group 6, 6A is clearly superior to 6B and 6D, whereas 6B and 6D do not differ from one another. F o r the backgrounds, 6A-6B-6D is the best, and 6A-6B is marginally better than 6B-6D. F o r b o t h groups o f chromosomes there are significant interactions between chromosomes and backgrounds (Table 3, chromosomes • b a c k g r o u n d s items). The best genotype in the trial, and, therefore, the one providing the most encouragement for exploiting the observed results, is N 1 A - T 1 D . Relationship with dosage The above analyses demonstrate that the r a n k order of chromosomes is 1D > 1 B > 1A, 6A > 6B = 6D, but do not directly address the effect o f varying the dosage of indi-
Table 2, Mean MST values for NTs, Ts and CSs in T88 Genetic background
Group 1 Chromosome
Background means
1A
1B
1B-ID
CS 5.79
N1A-T1B N1A-T1D 8.81 9.45
8.07
1A-ID
N1B-TIA 1.93
CS 6.26
NIB-T1D 6.58
4.92
N1D-T1A N1D-T1B 2.33 3.49
CS 6.13
3.98
T1A 5.43
T1B 6.93
T1D 8.27
6.88
Chromosome means
3.87
6.37
7.61
Genetic background
Group 6"
1A-1B 1A-1B-1D
1D
Chromosome
6B-6D
6A
6B
6D
Background means
CS 6.06
N6A-T6B 5.13
N6A-T6D 4.15
5.11
6A-6B
N6D-T6A N6D-T6B 6.08 4.90
CS 6.06
5.68
6A-6B-6D
T6A 7.08
T6B 6.06
T6D 6.15
6.43
Chromosome means
6.41
5.36
5.45
a Background 6A-6D omitted due to absence of N6B-T6D
Table 3. Analysis of variance testing differences between chromosomes and between backgrounds in T88" Source of variation
df
MS b
Replicate blocks Genotypes Group 1 chromosomes 1D vs (IA + 1B) 1A vs 1B Group 1 backgrounds IA-1B-ID vs (IB-1D + IA-1D + IA-1B) 1B-1D vs (1A-ID+ IA-1B) IA-1D vs 1A-IB Group I chromosomes • backgrounds Group 6 chromosomes Group 6 backgrounds Group 6 chromosomes x backgrounds Deviations Replicate blocks x genotypes interaction
4 37 2 1 1 3
0.12 Ns 13.09"** 72.52*** 82.42 *** 62.62*** 50.26***
1
17.17"**
1 126.97"** 1 6.63 *** 6 3.12"** 2 5.01"** 2 6.54*** 4 1.71"** 18 7.77*** 146 (2) 0.27
Standard error of difference between genotype means = 0.3258; between group I chromosomes=0.1629; between group 6 chromosomes and between backgrounds for both groups = 0.1881 b NS=P>0.05; ***=P
9 Springer-Verlag 1990
Dosage effects of chromosomes of homoeologous groups 1 and 6 upon bread-making quality in hexaploid wheat W.J. Rogers, J.M. Riekatson, E.J. Sayers and C.N. Law Cambridge Laboratory, JI Centre for Plant Science Research, Colney Lane, Norwich NR4 7UJ, UK Received January 12, 1990; Accepted February 23, 1990 Communicated by J.W. Snape
Summary. The endosperm storage proteins, glutenin and gliadin, are major determinants of bread-making quality in hexaploid wheat. Genes encoding them are located on chromosomes of homoeologous groups 1 and 6. Aneuploid lines of these groups in spring wheat cultivar 'Chinese Spring' have been used to investigate the effect of varying the dosage of chromosomes and chromosome arms upon bread-making quality, where quality has been assessed using the SDS-sedimentation test. Differences between the group 1 chromosomes for quality were greater than those between the group 6 chromosomes. The chromosomes were ranked within homoeologous groups for their effect on quality as follows ( > = b e t t e r quality): 1 D > I B > I A and 6 A > 6 B = 6 D . The relationship of chromosome dosage with quality was principally linear for four of the chromosomes, but not for 6B and 6D. Increases in the dosage of 1B, 6A and, especially, 1D, were associated with significant improvements in quality, whereas increases in the dosage of IA were associated with reductions in quality. The effects of 1A and 1D were such that the best genotype for quality was nullisomic 1A-tetrasomic 1D. For group 1, effects of the long arm appeared in general to be more important than effects of the short arm. For group 6, effects were found associated with the long arms as well as with the short arms, a surprising result in view of the absence of genes encoding storage proteins on the long arms. Significant interactions were found between chromosomes and genetic backgrounds, and between individual chromosomes. Analysis of trials grown over two years demonstrated that, although additive environmental differences over years and genotype x years interaction were present, they were relatively small in magnitude compared with purely genetic differences.
Key words: Triticum aestivum - Glutenin - Gliadin Aneuploid
Introduction Most studies of the relationship between endosperm storage proteins and bread-making quality in hexaploid wheat Triticum aestivum L. have sought to characterise the effect of allelic variation in genetic lines or commercial cultivars (Sozinov and Poperelya 1980; Payne et al. 1984, 1987a). The result has been the development of valuable methods for obtaining improvements in quality, by the identification of alleles associated with high quality, and their consequent assembly into favourable combinations during a breeding programme (Payne et al. 1984). Less attention has been directed towards the influence on quality of varying the dosage of genes known to affect quality. The principal series of homoeoloci encoding storage proteins that govern bread-making quality in hexaploid wheat are located on chromosomes comprising homoeologous groups 1 and 6 (Payne 1987, for review): Glu-1, located on the long arms of chromosomes of group 1, encoding high-molecular-weight (HMW) subunits of glutenin; Gli-1, located on the short arms of chromosomes of group 1, encoding co-gliadins, most of the 7gliadins and some of the fl-gliadins; Gli-2, located on the short arms of chromosomes of group 6, encoding egliadins, most of the fl-gliadins and some of the 7gliadins; and Glu-3, tightly linked to GIi-1 on group 1, encoding low-molecular-weight (LMW) subunits of glutenin. Several additional loci encoding endosperm proteins are located on the short arms of group 1, allelic variation at which has not yet been shown to influence quality. These loci include Gli-3, encoding o~-type gliadins and D subunits of LMW glutenin (Payne et al. 1988); loci encoding further minor classes of gliadins (Akhmedov and Metakovsky 1987); and Tri-l, encoding triticins that are classified as globulins (Singh et al. 1988) rather than as prolamins.
282 Studies of dosage have been primarily concerned with the effect of reduction in the dosage or, indeed, of complete removal, of particular loci, chromosome segments or whole chromosomes. F r o m such studies it has emerged that the following p h e n o m e n a detrimentally affect quality: (i) replacement of certain active alleles by null alleles, particularly at Glu-D1 (Payne et al. 1987b; Lawrence et al. 1988); (ii) replacement of part or all of the short arm of 1D and, to a lesser extent, of 1B, carrying Gli-1, by part of the short arms of homoeologous chromosomes from rye (Koebner et al. 1984) and Aegilops umbellulata (Harris 1983; Rogers et al. 1987); (iii) complete removal of the short arm of 1D in a ditelosomic line (Maystrenko etal. 1973); a n d (iv) reduction of the dosage of chromosome ID to the monosomic level (Welsh and H e h n 1964). Also, the replacement of Gli-B3 by a segment of chromosome 1U from Ae. umbeIlulata has been implicated in lowering quality (Harris 1983; Rogers et al. 1987). The current paper assesses the effect of varying the dosage of whole chromosomes and chromosome arms, both in the decreasing direction, i.e. below the disomic level, as in the previous studies cited above, a n d in the increasing direction, above the disomic level, in an attempt to identify beneficial effects on quality that might be open to exploitation via cytogenetic m a n i p u l a t i o n or molecular biology.
Materials and methods The study was principally based upon a trial of aneuploid lines of spring wheat cultivar 'Chinese Spring' (CS) grown during 1988 (trial T88). The CS aneuploids were originally produced by Dr. E. R. Sears, University of Missouri, and seed was kindly provided from stocks maintained by T. E. Miller and S. M. Reader at the Cambridge Laboratory. T88 consisted of 32 aneuploid lines involving chromosomes from homoeologous groups 1 and 6, as shown in Table 1, plus CS euploid controls. The abbreviations for the aneuploids given in Table I (NT, T, etc) will be used throughout the text. Six samples of CS were grown, where three were arbitrarily assigned to each group of aneuploids (1 and 6), enabling a number of a priori comparisons to be carried out during analysis (see 'Resuits'). A randomised complete block design was used, where each line was represented by a single plot of six plants in each of five independently randomised replicate blocks. Individual seeds were initially germinated on moist filter paper. Roots were taken from seedlings of monosomic families for cytological checking of Feulgen-stained mitotic preparations, and individuals with other than 41 chromosomes were discarded. Seedlings were planted into 4-cm pots and subsequently transplanted into soil in five unheated irrigated glasshouse compartments. Spacing was 15 cm between plants within a row (plot) and 20 cm between rows. Guard rows were sown at the ends of replicate blocks to prevent edge effects. During growth, only one fertiliser application was made, shortly after ear emergence, to guard against artifically boosting the protein content, which might obscure differences in quality due to differences in protein composition. Individual ears were bagged prior to anthesis to guar-
Table 1. Genotypes grown in trial T88 a Group 1 Nullisomic-tetrasomic (NT)
NtA-T1B NIA-TID
N1B-T1A NIB-T1D
N1D-T1A NID-T1B
Tetrasomic (T)
T1A
TIB
TID
Monosomic (M)
MIA
M1B
M1D
Ditelosomic (DT)
DT1AS DT1AL
DT1BS DTIBL
DT1DL
'Double ditelosomic' (DDT)
-
-
DDT1D b
Group 6 Nullisomic-tetrasomic (NT)
N6A-T6B N6A-T6D
N6B-T6A -
N6D-T6A N6D-T6B
Tetrasomic (T)
T6A
T6B
T6D
Monosomic (M)
M6A
M6B
M6D
Ditelosomic (DT)
-
DT6BS -
DT6DS DT6DL
Plus six independent euploid controls b This line is not a true double ditelosomic; one telosomic pair represents chromosome arm IDL. The other is not a true telosome; instead, it appears to be deleted long arm separated by the centromere from, presumably, a small segment of the short arm
antee self-pollination. Harvesting was carried out at maturity, and grain yield (GY) data was obtained for individual plants.
Quality tests Shortly after harvesting, seed was milled using a Tecator Cyclotec sample mill, and the flour was conditioned to 14%-15% moisture content. Grain protein contents (PC) based on 14% moisture content were determined using near infrared reflectance spectroscopy. Due to the very low GY from some of the aneuploids, it was necessary to apply a reliable quality test that could be performed on small amounts of material. For this purpose, a procedure was adapted from a test used to screen early generation durum wheat breeding lines, for gluten strength, using 1-g flour samples (Dick and Quick 1983), and termed the SDS-microsedimentation test (MST). Preliminary tests, in which the volume of a stock solution of SDS/lactic acid and weight of flour samples were varied, suggested that 0.7 g samples, 4 ml of HzO and 12 ml of stock solution would give the greatest differentiation between MST values for this material. The results were expressed as height in cm of sediment formed in a glass test tube (150 mm long x 16mm OD x 14ram ID) after 30min. Two replicate MST determinations were made on each flour sample.
Results
Correlation between characters N o correlation was found between PC a n d MST, suggesting that overall differences between genotypes for M S T were due to qualitative differences in protein composition rather than to quantitative differences in protein
283 amount. To ensure against misinterpretation, M S T was analysed b o t h with and without PC as a covariate, and the conclusions differed appreciably in only one case (considered below). In general, only analyses unadjusted for PC are reported. Individual chromosome effects The mean performance o f each genotype is given in Table 2. F o r analysis, the NTs and Ts of each homoeologous group, plus three o f the CS euploids, provide a balanced and orthogonal set o f genotypes for identifying the effects of individual chromosomes and combinations of chromosomes, in an extension of the design of Pink and Law (1985). In theory, with the families arranged as shown in Table 2, each c h r o m o s o m e appears in four ' b a c k g r o u n d s ' , i.e. the three pairwise combinations of the remaining two chromosomes in the relevant homoeologous group, plus the combination of all three chromosomes. F o r each chromosome, the mean over the four b a c k g r o u n d s can be c o m p a r e d to determine which chrom o s o m e is best for quality; similarly, the means of the b a c k g r o u n d s can be compared. Interactions between chromosomes can also be tested. In practice, b a c k g r o u n d 6A-6D has h a d to be excluded due to the enforced absence of N 6 B - T 6 D from the trial (Table 1). A n analysis o f variance of the d a t a is given in Table 3, including orthogonal comparisons between chromosomes. Previous evidence suggested that valid comparisons would be 1D against the mean o f 1A and 1B, and, given that 1A carries a null allele at Glu-A1, IB-1D against the mean o f 1A-1B and 1A-1D. Also the euploid b a c k g r o u n d can be c o m p a r e d with the m e a n o f the three pairwise comparisons 1A-1B, 1A-1D and 1B-1D. N o a priori comparisons have been calculated for group 6, but differences can be c o m p a r e d using the a p p r o p r i a t e stand a r d error of the difference between means. The analyses clearly demonstrate the superiority of 1D over the mean of 1A and 1B, and o f 1B over I A . The best b a c k g r o u n d is 1B-1D, and 1A-1D is superior to 1A-lB. F o r group 6, 6A is clearly superior to 6B and 6D, whereas 6B and 6D do not differ from one another. F o r the backgrounds, 6A-6B-6D is the best, and 6A-6B is marginally better than 6B-6D. F o r b o t h groups o f chromosomes there are significant interactions between chromosomes and backgrounds (Table 3, chromosomes • b a c k g r o u n d s items). The best genotype in the trial, and, therefore, the one providing the most encouragement for exploiting the observed results, is N 1 A - T 1 D . Relationship with dosage The above analyses demonstrate that the r a n k order of chromosomes is 1D > 1 B > 1A, 6A > 6B = 6D, but do not directly address the effect o f varying the dosage of indi-
Table 2, Mean MST values for NTs, Ts and CSs in T88 Genetic background
Group 1 Chromosome
Background means
1A
1B
1B-ID
CS 5.79
N1A-T1B N1A-T1D 8.81 9.45
8.07
1A-ID
N1B-TIA 1.93
CS 6.26
NIB-T1D 6.58
4.92
N1D-T1A N1D-T1B 2.33 3.49
CS 6.13
3.98
T1A 5.43
T1B 6.93
T1D 8.27
6.88
Chromosome means
3.87
6.37
7.61
Genetic background
Group 6"
1A-1B 1A-1B-1D
1D
Chromosome
6B-6D
6A
6B
6D
Background means
CS 6.06
N6A-T6B 5.13
N6A-T6D 4.15
5.11
6A-6B
N6D-T6A N6D-T6B 6.08 4.90
CS 6.06
5.68
6A-6B-6D
T6A 7.08
T6B 6.06
T6D 6.15
6.43
Chromosome means
6.41
5.36
5.45
a Background 6A-6D omitted due to absence of N6B-T6D
Table 3. Analysis of variance testing differences between chromosomes and between backgrounds in T88" Source of variation
df
MS b
Replicate blocks Genotypes Group 1 chromosomes 1D vs (IA + 1B) 1A vs 1B Group 1 backgrounds IA-1B-ID vs (IB-1D + IA-1D + IA-1B) 1B-1D vs (1A-ID+ IA-1B) IA-1D vs 1A-IB Group I chromosomes • backgrounds Group 6 chromosomes Group 6 backgrounds Group 6 chromosomes x backgrounds Deviations Replicate blocks x genotypes interaction
4 37 2 1 1 3
0.12 Ns 13.09"** 72.52*** 82.42 *** 62.62*** 50.26***
1
17.17"**
1 126.97"** 1 6.63 *** 6 3.12"** 2 5.01"** 2 6.54*** 4 1.71"** 18 7.77*** 146 (2) 0.27
Standard error of difference between genotype means = 0.3258; between group I chromosomes=0.1629; between group 6 chromosomes and between backgrounds for both groups = 0.1881 b NS=P>0.05; ***=P
